MR imaging of the liver on multi-channel MR systems at 1.5 and 3Tesla – Initial experiences

Purpose: Due to the higher signal-to-noise ratio (SNR) MR imaging at 3Tesla (T) systems offers new perspectives; however, there are a number of challenges for abdominal MRI at 3T despite the obvious advantage of a higher SNR, including a fourfold higher specific absorption rate as well as the presence of dielectric artifacts. The objective of this study was to compare image quality and diagnostic accuracy of liver MRI with liver-specific contrast agents at 1.5T and 3T using multi-channel MR-systems.

Methods: 26 patients which were referred for clinical reasons to a MR examination of the liver with the approved liver-specific contrast agent Primovist® (Schering AG, Berlin) were evaluated. MR examinations were performed on 1.5T (Magnetom Avanto) and 3T (Magnetom Tim Trio) 32-channel MR systems with parallel imaging (Siemens Medical Solutions, Erlangen). The sequences were adjusted according to the specific requirements of the corresponding field strength. The spatial resolution on the 3T system was increased up to 45% by increasing the in-plane matrix and decreasing the slice thickness (for example for the T1-w 3D GRE sequence: 1.5T: 4mm slice thickness (sl), Matrix 256×180; TR 5.16/TE 1.93/FA 15, 19 seconds breath-hold; 3T: 2mm sl, matrix 320×260, TR 4.36/TE 1.59/FA 15, 17 seconds breath-hold). Patients were randomized with regard to field strength of the pre-contrast examination and the order of the double examination in the liver-specific phase (1.5T first and then 3T or vice versa). In a blinded reading two experienced readers evaluated image quality and delineation of lesions on a 5-point scale (1=not interpretable; 5=excellent image quality/delineation). Moreover, the frequency of specific artifacts was evaluated.

Results: Image quality was very good to excellent at 1.5 and 3T and did not show significant differences for both readers. (Reader 1 (1.5T vs. 3T), right liver lobe (LL): 4.2 vs. 4.2; left LL: 4.2 vs. 3.9; Reader 2, right LL: 4.2 vs. 4.2; left LL: 4.3 vs. 4.0). Delineation of lesions was also not significantly different for 1.5 and 3T. With regard to specific artifacts at 1.5 Tesla reader 1 noted 0 cases of dielectric artifacts, 0 cases of inhomogeneous fat-sat, 6 cases of susceptibility artifacts and 11 cases of motion artifacts. The same evaluation for 3T showed dielectric artifacts in 9, inhomogeneous fat-sat in 5, susceptibility artifacts in 7 and motion artifacts in 10 cases.

Conclusion: First results on multi-channel MR systems show that MR examinations of the liver are feasible with high quality at 1.5T as well as 3T. High-field specific artifacts as dielectric artifacts or inhomogeneous fat-saturation were more frequent at 3T; however, image quality in general and delineation of lesions were not negatively affected. Further studies have to investigate whether the possibility to increase the spatial resolution at 3T can lead to an increased detection rate for focal liver lesions.